Regional and Seasonal Trends in Park Visitation
Data Importing and Cleaning
res=GET('https://developer.nps.gov/api/v1/parks?limit=500&api_key=B9nDpbkbrb3kSOjz6kXSxMJ3d6MSpUvt1QqYdeyn')
data = res %>% content("text") %>% jsonlite::fromJSON() %>% as_tibble()
NPS_data=data %>% unnest(data) %>% select(fullName,latitude,longitude,topics, activities,states, parkCode) %>% janitor::clean_names() %>%
mutate(
latitude = as.numeric(latitude),
longitude = as.numeric(longitude)
) %>% unnest(activities, names_sep = "_") %>%
unnest(topics, names_sep = "_")
visitation_data <-
read_csv("data/Query Builder for Public Use Statistics (1979 - Last Calendar Year).csv") %>%
janitor::clean_names() %>%
mutate(unit_code = tolower(unit_code)) %>%
rename(park_code = unit_code,
full_name = park_name) %>%
select(full_name, park_code, park_type, region, state, year, month, recreation_visits, tent_campers, rv_campers, tent_campers, backcountry)
combined_data <- full_join(NPS_data, visitation_data, by = c("park_code"))Trends across Time
Total Trend by Park Type Across Months
visitation_data %>%
group_by(park_type, month) %>%
summarize(total_visitation = sum(recreation_visits)) %>%
plot_ly(x = ~as.factor(month), y = ~total_visitation, type = "scatter", mode = "lines", color = ~park_type)visitation_data %>%
group_by(park_type) %>%
summarize(parks = n_distinct(full_name)) %>%
arrange(desc(parks)) %>%
knitr::kable()| park_type | parks |
|---|---|
| National Monument | 73 |
| National Historic Site | 72 |
| National Historical Park | 58 |
| National Park | 55 |
| National Memorial | 32 |
| National Recreation Area | 18 |
| Park (Other) | 11 |
| National Battlefield | 10 |
| National Seashore | 10 |
| National Military Park | 9 |
| National Preserve | 7 |
| National Wild & Scenic River | 7 |
| National Battlefield Park | 4 |
| National Parkway | 4 |
| National River | 4 |
| National Lakeshore | 3 |
| International Historic Site | 1 |
| National Reserve | 1 |
As seen in the graph, national parks have the highest total visitation among all park types. Interestingly, there more national monuments, national historic sites, and national historic parks than national parks, even though all of these have less total visitation than national parks.
Total Visitation by Visitation Type and Season
The visitation_data data set also includes information
on visitation split by tent_campers,
backcountry, and rv_campers. We were
interested in looking at the total visitation split by these three
visitation types and by season.
visitation_data %>%
mutate(season = case_when(
month %in% c(12,1,2) ~ "Winter",
month %in% c(3,4,5) ~ "Spring",
month %in% c(6,7,8) ~ "Summer",
TRUE ~ "Fall"
)) %>%
group_by(season) %>%
summarize(total_tent = sum(tent_campers),
backcountry_visits = sum(backcountry),
total_rv = sum(rv_campers)) %>%
pivot_longer(
total_tent:total_rv,
values_to = "total_visit",
names_to = "type_visit",
names_prefix = "total_"
) %>%
plot_ly(x = ~season, y = ~total_visit, color = ~type_visit, type = "bar")In fall, winter, and spring, the highest visitation is
tent_campers, with a peak summer. Interestingly and
unsurprisingly, the highest visitation in winter is
rv_campers. This is probably due to weather conditions not
permitting tent camping, but allows for rv camping. For fall, winter,
and spring there is a similar trend where backcountry has
the lowest total visitation and tent_campers has the
highest visitation. In winter, rv_campers has the highest
total visitation, then tent_campers, and then
backcountry.
Average Park Visitation by Season and Park Type
Next, we were interested in seeing the trends in park visitation by season and park type.
park_types <- unique(visitation_data$park_type)
plots <- list()
for (i in seq_along(park_types)) {
plots[[i]] <- visitation_data %>%
filter(park_type == park_types[i]) %>%
mutate(season = case_when(
month %in% c(12, 1, 2) ~ "Winter",
month %in% c(3, 4, 5) ~ "Spring",
month %in% c(6, 7, 8) ~ "Summer",
TRUE ~ "Fall"
)) %>%
group_by(season) %>%
summarize(avg_visits = mean(recreation_visits, na.rm = TRUE)) %>%
ggplot(aes(x = season, y = avg_visits, fill = season)) +
geom_col() +
ggtitle(paste( park_types[i]))
} Most of the parks have a similar trend in average visitation, where summer is the highest and winter is the lowest.
National Lakeshore Seasonal Trends and Activities
For national lakeshore parks, there is a really stark
difference in average visits between summer and the other seasons.
activities_count <-
combined_data %>%
filter(park_type == "National Lakeshore") %>%
distinct(activities_name) %>%
nrow()
combined_data %>%
filter(park_type == "National Lakeshore") %>%
distinct(activities_name) %>%
head() %>% knitr::kable()| activities_name |
|---|
| Boating |
| Motorized Boating |
| Sailing |
| Boat Tour |
| Camping |
| Backcountry Camping |
Based on the activities offered in National Lakeshore, we can see
that most of the activities revolve around water, such as
boating, sailing, and boat tours.
Interestingly, at National Lakeshore, there are skiing activities, but
the winter average visits are still the lowest.
National Preserve Seasonal Trends and Activities
plots[10]## [[1]]
Another interesting pattern is that in National Preserve, there is a higher average visitation in spring than summer and winter has a higher visitation than fall.
activities_count <-
combined_data %>%
filter(park_type == "National Preserve") %>% distinct(activities_name) %>% nrow()
combined_data %>%
filter(park_type == "National Preserve") %>% distinct(activities_name) %>%
head() %>%
knitr::kable()| activities_name |
|---|
| Arts and Culture |
| Cultural Demonstrations |
| Auto and ATV |
| ATV Off-Roading |
| Scenic Driving |
| Astronomy |
There are 63 activities offered at this park. It is near the middle of the distribution of total activities, so the trend is not due to having many more activities offered than other park types. It seems like there is a much bigger variety in the types of activities offered at these parks. There are tours, stargazing, biking, picnicking, hunting, water activities, wildlife watching, museums and stores, as well as winter activities like skiing, and snowshoeing.
International Historic Site Seasonal Trends and Activities
plots[18]## [[1]]
This plot is interesting because there is no average visitation in winter.
activities_count <-
combined_data %>%
filter(park_type == "International Historic Site") %>% distinct(activities_name) %>% nrow()
combined_data %>%
filter(park_type == "International Historic Site") %>% distinct(activities_name) %>%
head() %>%
knitr::kable()| activities_name |
|---|
| Paddling |
| Junior Ranger Program |
| Wildlife Watching |
There are 3 activities offered at this park type. It makes sense that
there is no visitation in the winter because the only activities offered
are paddling, junior ranger program and
wildlife watching. There are no activities that would
really be done in the winter at these parks.
Average Visitation by Visit and Park Type
park_types <- unique(visitation_data$park_type)
plots <- list()
for (i in seq_along(park_types)) {
visit_summary <- visitation_data %>%
drop_na() %>%
filter(park_type == park_types[i]) %>%
mutate(season = case_when(
month %in% c(12, 1, 2) ~ "Winter",
month %in% c(3, 4, 5) ~ "Spring",
month %in% c(6, 7, 8) ~ "Summer",
TRUE ~ "Fall"
)) %>%
group_by(season) %>%
summarize(
mean_tent = mean(tent_campers, na.rm = TRUE),
mean_backcountry = mean(backcountry, na.rm = TRUE),
mean_rv = mean(rv_campers, na.rm = TRUE)
) %>%
pivot_longer(
cols = starts_with("mean_"),
names_to = "type_visit",
values_to = "mean",
names_prefix = "mean_"
)
plots[[i]] <- ggplot(visit_summary, aes(x = season, y = mean)) +
geom_col() +
facet_grid(~type_visit) +
ggtitle(paste("Average Visits for", park_types[i])) +
theme(axis.text.x = element_text(angle = 90, vjust = 0.5, hjust = 1))
} No Visitation
plots[[4]] + plots[[8]] + plots[[17]] + plots[[18]]
National Historic Sites, National Memorials, National Battlefield
Parks, International Historic Sites all had no faceted average visits.
These four park types do not have any rv_campers,
backcountry or tent_campers.
Only Backcountry Visitation
plots[[16]] + plots[[15]] + plots[[6]]
National Reserve only had backcountry visitation in
spring. National Military Park had backcountry visitation
in all seasons, with mean visitation being highest in spring.
National Battlefield had backcountry visitation in all
seasons, with mean visitation being lowest in summer and highest in
spring.
No RV Visitation
plots[13]## [[1]]
National Wild and Scienic River has backcountry and
tent visitation, no rv visitation.
Interestingly though, the mean tent visitation is low for
all seasons, with a slight peak in summer.
Park (Other) Visitation
plots[14]## [[1]]
This plot was interesting because the park type
Park (Other) has high tent mean visitation,
rv visitation, but low backcountry visitation.
It is the highest in spring and summer.
map of all national parks in US
regional_data <- NPS_data %>%
mutate(region = case_when(
states %in% c("CT", "RI", "NH", "VT", "NJ", "NY", "PA", "MD", "ME", "MA") ~ "northeast",
states %in% c("IL","IN", "MI", "OH", "WI", "IA", "KS", "MN", "MO", "NE", "ND", "SD") ~ "midwest",
states %in% c("FL", "GA", "NC", "SC", "VA", "DE", "WV", "AL", "KY", "MS", "TN", "AR", "LA", "OK", "TX", "DC") ~ "south",
states %in% c("AK", "CA", "HI", "OR", "WA", "AZ", "CO", "ID", "MT", "NV", "NM", "UT", "WY") ~ "west",
states %in% c("VI", "AS", "GU", "PR") ~ "u.s. territory",
TRUE ~ "no state data"
))
g <- list(
scope = 'usa',
projection = list(type = 'albers usa'),
showland = TRUE,
landcolor = toRGB("gray95"),
subunitcolor = toRGB("gray85"),
countrycolor = toRGB("gray85"),
countrywidth = 0.5,
subunitwidth = 0.5
)
fig <- plot_geo(regional_data, lat = ~latitude, lon = ~longitude)
fig <- fig %>% add_markers(
text = ~paste(full_name, states, sep = "<br />"),
color = ~region, symbol = I("circle"), size = I(8), hoverinfo = "text"
)
fig <- fig %>% layout(
title = 'US National Parks', geo = g
)
fig